Artificial Intelligence: Machine Learning for Early Detection of Pests and Diseases in Basic Crops, Nicaragua

The present article highlights relevant aspects of the development process of the mobile application that incorporates Machine Learning techniques to early detect pests and diseases in staple grain crops such as corn, beans, and sorghum, which are essential for human consumption in Nicaragua. Agile development methodology Scrum was used, technologies such as Android Studio, Java programming language, Google Teachable Machine for training the machine learning model, and TensorFlow Lite for incorporating the model into the mobile application were adopted. The results show a Sprint with its user stories, which were turned into functionalities that include the model for image recognition with an accuracy of 95.8% using a dataset of 252 images of healthy and diseased crops. The methodology indicates the organization of programming according to the Model-View-Controller pattern and the metrics used by the model. The conclusions emphasize details of the results obtained in Sprint#1. In the end, challenges to overcome in applying machine learning in the agricultural sector are also mentioned.

• Egile nagusia: Urbina Cienfuegos, Saira María (author)
• Beste egile batzuk: Bravo Rivas, Jazcar Josué (author)
• Formatua: article
• Hizkuntza: spa
• Argitaratua: 2025
• Gaiak: machine learning; plagas; enfermedades; cultivos; pests; diseases; crops
• Sarrera elektronikoa: https://revistadigital.uce.edu.ec/index.php/INGENIO/article/view/7221